rRNADepletion_s4ter

Setup

library(tidyverse)
library(magrittr)
library(parallel)
library(ngsReports)
library(here)
library(scales)
library(ggpubr)
library(kableExtra)
library(AnnotationHub)
library(ensembldb)
library(edgeR)
library(corrplot)
library(DT)
library(ggrepel)

if (interactive()) setwd(here::here())
theme_set(theme_bw())
cores <- detectCores() - 1

Sequence information

ah_Dr <- AnnotationHub() %>%
  subset(species == "Danio rerio") %>%
  subset(rdataclass == "EnsDb")
ensDb_Dr <- ah_Dr[["AH83189"]]
trEns_Dr <- transcripts(ensDb_Dr) %>%
  mcols() %>% 
  as_tibble()
trLen_Dr <- exonsBy(ensDb_Dr, "tx") %>%
  width() %>%
  vapply(sum, integer(1))
geneGcLen_Dr <- trLen_Dr %>%
  enframe() %>%
  set_colnames(c("tx_id", "length")) %>%
  left_join(trEns_Dr) %>%
  group_by(gene_id) %>% 
  summarise(
    aveLen = mean(length),
    maxLen = max(length), 
    aveGc = sum(length * gc_content) / sum(length),
    longestGc = gc_content[which.max(length)[[1]]]
  ) %>%
  mutate(
    aveGc =  aveGc / 100,
    longestGc = longestGc / 100
  )
trGcLen_Dr <- trLen_Dr %>%
  enframe() %>%
  set_colnames(c("tx_id", "length")) %>%
  left_join(trEns_Dr) %>%
  group_by(tx_id) %>% 
  summarise(
    aveLen = mean(length),
    maxLen = max(length), 
    aveGc = sum(length * gc_content) / sum(length),
    longestGc = gc_content[which.max(length)[[1]]]
  ) %>%
  mutate(
    aveGc =  aveGc / 100,
    longestGc = longestGc / 100
  )
genesGR_Dr <- genes(ensDb_Dr)
mcols(genesGR_Dr) <- mcols(genesGR_Dr)[c("gene_id", "gene_name", 
                                         "gene_biotype", "entrezid")]
txGR_Dr <- transcripts(ensDb_Dr)
mcols(txGR_Dr) <- mcols(txGR_Dr)[c("tx_id", "tx_name", 
                                   "tx_biotype", "tx_id_version", "gene_id")]

An EnsDb object was obtained for Ensembl release 101 using the AnnotationHub package. This provided the GC content and length for every gene and transcript in the release. For zebrafish, this consists of 37241 genes and 65905 transcripts.

Raw data

This is a total RNA-seq dataset generated from a 3-way comparison of WT zebrafish (Danio rerio) with heterozygous mutants (psen2^S4Ter/+) and homozygous mutants (psen2^S4Ter/S4Ter). A previous analysis of this dataset identified the possibility of incomplete ribosomal RNA (rRNA) removal. The following analysis involves an investigation into possible reasons for incomplete rRNA removal and any bias this introduces into the data.

Sample information

files <- list.files(
  path = "/hpcfs/users/a1647910/20200310_rRNADepletion/1_Psen2S4Ter/0_rawData/FastQC",
  pattern = "zip",
  full.names = TRUE
)

samples <- tibble(
  sample = str_remove(basename(files), "_fastqc.zip"),
  dataset = NA,
  organism = NA
) %>%
  mutate(
    dataset = ifelse(
      str_detect(sample, "Ps2Ex"), "Psen2S4Ter", dataset
    ),
    organism = ifelse(
      str_detect(sample, "Ps2Ex"), "zebrafish", organism
    )
  )
datasets <- samples$dataset %>% 
  unique()

The following analysis involves 12 paired-end samples across 1 dataset(s): Psen2S4Ter.

Library sizes

rawFqc <- files %>%
  FastqcDataList()

data <- grep("Ps2Ex", fqName(rawFqc))
labels <- rawFqc[data] %>%
  fqName() %>%
  str_remove("_6month_07_07_2016_F3") %>%
  str_remove("\\.fastq\\.gz") %>%
  str_remove("Ps2Ex3M1_")
rawLib <- plotReadTotals(rawFqc[data]) +
  labs(subtitle = "Psen2S4Ter") + 
  scale_x_discrete(labels = labels)

The library sizes of the unprocessed dataset(s) range between 27,979,654 and 37,144,975 reads.

rawLib

GC content

rRNA transcripts are known to have high GC content. Therefore, inspecting the GC content of the raw reads is a logical start point for detecting incomplete rRNA removal. A spike in GC content at > 70% is expected if this is the case.

plotly::ggplotly(
  plotGcContent(
    x = rawFqc[data], 
    plotType = "line",
    gcType = "Transcriptome",
    species = "Drerio"
  ) +
    labs(title = "Psen2S4Ter Dataset (D. rerio)") + 
    theme(legend.position="none")
) 

GC content of reads in the dataset. Clear spikes above 70% GC are observed, which is likely due to incomplete rRNA depletion.

Overrepresented sequences

The top 30 overrepresented sequences were analysed using blastn and were found to be predominantly rRNA sequences.

getModule(rawFqc, "Overrep") %>% 
  group_by(Sequence, Possible_Source) %>% 
  summarise(`Found In` = n(), `Highest Percentage` = max(Percentage)) %>% 
  arrange(desc(`Highest Percentage`), desc(`Found In`)) %>% 
  ungroup() %>% 
  dplyr::slice(1:30) %>%
  mutate(`Highest Percentage` = percent_format(0.01)(`Highest Percentage`/100)) %>%
  kable(
    align = "llrr", 
    caption = paste(
      "Top", nrow(.),"Overrepresented sequences.",
      "The number of samples they were found in is shown,",
      "along with the percentage of the most 'contaminated' sample."
    )
  ) %>%
  kable_styling(
    bootstrap_options = c("striped", "hover", "condensed", "responsive")
  )

Top 30 Overrepresented sequences. The number of samples they were found in is shown, along with the percentage of the most ‘contaminated’ sample.

Sequence	Possible_Source	Found In	Highest Percentage
GTGGGTTCAGGTAATTAATTTAAAGCTACTTTCGTGTTTGGGCCTCTAGC	No Hit	12	1.72%
CTGGGGGAGCGGCCGCCCCGCGGCGCCCCCTCTCGTTCCCGTCTCCGGAG	No Hit	10	1.69%
CCGCTGTATTACTCAGGCTGCACTGCAGTGTCTATTCACAGGCGCGATCC	No Hit	12	1.32%
GGCCCGGCGCACGTCCAGAGTCGCCGCCGCACACCGCAGCGCATCCCCCC	No Hit	9	1.31%
CTCCTGAAAAGGTTGTATCCTTTGTTAAAGGGGCTGTACCCTCTTTAACT	No Hit	11	1.11%
GGTTCAGGTAATTAATTTAAAGCTACTTTCGTGTTTGGGCCTCTAGCATC	No Hit	12	1.09%
GGGGTGTACGAAGCTGAACTTTTATTCATCTCCCAGACAACCAGCTATTG	No Hit	12	1.07%
GGCCCGGCGCACGTCCAGAGTCGCCGCCGCGCACCGCAGCGCATCCCCCC	No Hit	10	1.03%
CGAGAGGCTCTAGTTGATATACTACGGCGTAAAGGGTGGTTAAGGAACAA	No Hit	12	1.01%
GGGGGAGCGGCCGCCCCGCGGCGCCCCCTCTCGTTCCCGTCTCCGGAGCG	No Hit	9	0.87%
CCTCCTTCAAGTATTGTTTCATGTTACATTTTCGTATATTCTGGGGTAGA	No Hit	12	0.82%
CCCGCTGTATTACTCAGGCTGCACTGCAGTGTCTATTCACAGGCGCGATC	No Hit	12	0.79%
GTTCAGGTAATTAATTTAAAGCTACTTTCGTGTTTGGGCCTCTAGCATCT	No Hit	12	0.79%
GGGTTCAGGTAATTAATTTAAAGCTACTTTCGTGTTTGGGCCTCTAGCAT	No Hit	12	0.78%
CGGGTCGGGTGGGTGGCCGGCATCACCGCGGACCTCGGGCGCCCTTTTGG	No Hit	12	0.73%
GGGCCTCTAGCATCTAAAAGCGTATAACAGTTAAAGGGCCGTTTGGCTTT	No Hit	11	0.68%
CAGTGGCGTGCGCCTGTAATCCAAGCTACTGGGAGGCTGAGGCTGGCGGA	No Hit	11	0.64%
CGGGTCGGGTGGGTAGCCGGCATCACCGCGGACCTCGGGCGCCCTTTTGG	No Hit	12	0.57%
CTTAGACGACCTGGTAGTCCAAGGCTCCCCCAGGAGCACCATATCGATAC	No Hit	11	0.54%
AGCTGGGGAGATCCGCGAGAAGGGCCCGGCGCACGTCCAGAGTCGCCGCC	No Hit	11	0.53%
GGCCTCTAGCATCTAAAAGCGTATAACAGTTAAAGGGCCGTTTGGCTTTA	No Hit	10	0.53%
CAGCCTATTTAACTTAGGGCCAACCCGTCTCTGTGGCAATAGAGTGGGAA	No Hit	12	0.51%
GGGTGGGTGGCCGGCATCACCGCGGACCTCGGGCGCCCTTTTGGACGTGG	No Hit	10	0.50%
GGGAGCGGCCGCCCCGCGGCGCCCCCTCTCGTTCCCGTCTCCGGAGCGCG	No Hit	9	0.49%
CTGGGAGATGAATAAAAGTTCAGCTTCGTACACCCCAAATTAAAAAATTA	No Hit	10	0.48%
GCCTATTTAACTTAGGGCCAACCCGTCTCTGTGGCAATAGAGTGGGAAGA	No Hit	12	0.47%
GGTCGGGTGGGTGGCCGGCATCACCGCGGACCTCGGGCGCCCTTTTGGAC	No Hit	11	0.45%
CCCCCGAACCCTTCCAAGCCGAACCGGAGCCGGTCGCGGCGCACCGCCGA	No Hit	10	0.45%
GTCGGGTGGGTGGCCGGCATCACCGCGGACCTCGGGCGCCCTTTTGGACG	No Hit	10	0.43%
GCCCACTACGACAACGTGTTTTGTAAATTATGATCTTTATTCTCCTGAAA	No Hit	10	0.43%

Trimmed data

Raw libraries were trimmed using cutadapt v1.14 to remove Illumina adapter sequences. Bases with PHRED score < 30, NextSeq-induced polyG runs and reads shorter than 35bp were also removed.

trimFqc <- list.files(
  path = "/hpcfs/users/a1647910/20200310_rRNADepletion/1_Psen2S4Ter/1_trimmedData/FastQC",
  pattern = "zip",
  full.names = TRUE
) %>%
  FastqcDataList()
trimStats <- readTotals(rawFqc) %>%
  dplyr::rename(Raw = Total_Sequences) %>%
  left_join(readTotals(trimFqc), by = "Filename") %>%
  dplyr::rename(Trimmed = Total_Sequences) %>%
  mutate(
    Discarded = 1 - Trimmed/Raw,
    Retained = Trimmed / Raw
  )

After trimming of adapters between 4.16% and 5.08% of reads were discarded.

Aligned data

Trimmed reads were:

1. Aligned to rRNA sequences using the BWA-MEM algorithm to estimate the proportion of reads that were of rRNA origin within each sample. BWA-MEM is recommended for high-quality queries of reads ranging from 70bp to 1Mbp as it is faster and more accurate that alternative algorithms BWA-backtrack and BWA-SW.

2. Aligned to the Danio rerio GRCz11 genome (Ensembl release 101) using STAR v2.7.0d and summarised with featureCounts from the Subread v1.5.2 package. These counts were used for all gene-level analysis.

rRNA proportions

rRnaProp <- read.delim(
  "/hpcfs/users/a1647910/20200310_rRNADepletion/1_Psen2S4Ter/3_bwa/log/samples.mapped.all", 
  sep = ":", 
  col.names = c("sample", "proportion"), 
  header = FALSE
) %>% 
  mutate(
    sample = str_remove_all(sample, "_6month_F3|[0-9]*_Ps2Ex3M1_|.mapped"),
    sample = basename(sample),
    proportion = proportion/100,
    dataset = "Psen2S4Ter",
    organism = "zebrafish",
    group = str_extract(sample, "(WT|Heter|Hom)")
  ) %>%
  as_tibble()
rRnaProp$dataset %<>%
  factor(levels = c("Psen2S4Ter"))

rRnaProp %>%
  ggplot(aes(sample, proportion)) +
  geom_bar(stat = "identity", position = "dodge") +
  facet_wrap(~dataset, scales = "free_x") +
  scale_y_continuous(labels = percent) +
  labs(x = "Sample", y = "Percent of Total", fill = "Read pair") +
  theme(axis.text.x = element_text(angle = 90, hjust = 1))

Percentages of each library that align to rRNA sequences with bwa mem.

rRnaProp %>%
  ggplot(aes(group, proportion, fill = group)) +
  geom_boxplot() +
  scale_y_continuous(labels = percent) +
  labs(x = "Genotype", y = "Percent of total RNA", title = "rRNA proportions of each genotype") +
  scale_fill_discrete(
    name = "Genotype"
  )

Gene GC content and length

dgeList <- read_tsv("/hpcfs/users/a1647910/20200310_rRNADepletion/1_Psen2S4Ter/4_star2pass/featureCounts/genes.out") %>%
  set_colnames(basename(colnames(.))) %>%
  set_colnames(str_remove(colnames(.), "Aligned.sortedByCoord.out.bam")) %>%
  set_colnames(str_remove(colnames(.), "_6month_F3")) %>%
  set_colnames(str_remove(colnames(.), "[0-9]*_Ps2Ex3M1_")) %>%
  as.data.frame() %>%
  column_to_rownames("Geneid") %>%
  DGEList() %>%
  calcNormFactors()

dgeList$genes <- genesGR_Dr[rownames(dgeList),]
mcols(dgeList$genes) %<>% 
  as.data.frame() %>% 
  left_join(geneGcLen_Dr)
addInfo <- tibble(
  sample = rRnaProp$sample,
  dataset = "Psen2S4Ter",
  organism = "zebrafish",
  rRNA = rRnaProp$proportion
)
dgeList$samples %<>%
  rownames_to_column("rowname") %>%
  mutate(sample = rowname) %>%
  left_join(addInfo) %>%
  column_to_rownames("rowname")
dgeList$samples$filenames <- list.files(
  "/hpcfs/users/a1647910/20200310_rRNADepletion/1_Psen2S4Ter/2_alignedData/bam", 
  pattern = ".bam$", 
  full.names = TRUE
)
dgeList$samples$group <- colnames(dgeList) %>%
  str_extract("(WT|Heter|Hom)") %>%
  factor(levels = c("WT", "Heter", "Hom"))

gcInfo <- function(x) {
  x$counts %>%
    as.data.frame() %>%
    rownames_to_column("gene_id") %>%
    as_tibble() %>%
    pivot_longer(
      cols = colnames(x), 
      names_to = "sample", 
      values_to = "counts"
    ) %>%
    dplyr::filter(
      counts > 0
    ) %>%
    left_join(
      geneGcLen_Dr
    ) %>%
    dplyr::select(
      ends_with("id"), sample, counts, aveGc, maxLen
    ) %>%
    split(f = .$sample) %>%
    lapply(
      function(x){
        DataFrame(
          gc = Rle(x$aveGc, x$counts),
          logLen = Rle(log10(x$maxLen), x$counts)
        )
      }
    ) 
}
gcSummary <- function(x) {
  x %>%
    vapply(function(x){
      c(mean(x$gc), sd(x$gc), mean(x$logLen), sd(x$logLen))
    }, numeric(4)
    ) %>%
    t() %>%
    set_colnames(
      c("mn_gc", "sd_gc", "mn_logLen", "sd_logLen")
    ) %>%
    as.data.frame() %>%
    rownames_to_column("sample") %>%
    as_tibble()
}

rle <- gcInfo(dgeList)

sumGc <- gcSummary(rle)

a <- sumGc %>%
  left_join(dgeList$samples) %>%
  ggplot(aes(rRNA, mn_logLen)) +
  geom_point(aes(colour = group), size = 3) +
  geom_smooth(method = "lm") +
  scale_x_continuous(labels = percent) +
  labs(
    x = "rRNA Proportion of Initial Library",
    y = "Mean log(Length)",
    colour = "Genotype"
  ) 
b <- sumGc %>%
  left_join(dgeList$samples) %>%
  ggplot(aes(rRNA, mn_gc)) +
  geom_point(aes(colour = group), size = 3) +
  geom_smooth(method = "lm") +
  scale_y_continuous(labels = percent) +
  scale_x_continuous(labels = percent) +
  labs(
    x = "rRNA Proportion of Initial Library",
    y = "Mean GC Content",
    colour = "Genotype"
  )
ggarrange(
  a, b, ncol = 2, nrow = 1, 
  common.legend = TRUE, legend = "bottom"
) %>%
  annotate_figure("PsenS4Ter Dataset (D. rerio)")

Comparison of residual bias potentially introduced by incomplete rRNA removal. Regression lines are shown along with standard error bands for each comparison.

PCA

genes2keep <- dgeList %>%
  cpm() %>%
  is_greater_than(1) %>%
  rowSums() %>%
  is_weakly_greater_than(6)
dgeFilt <- dgeList[genes2keep,, keep.lib.sizes = FALSE] %>%
  calcNormFactors()

pca <- cpm(dgeFilt, log = TRUE) %>%
  t() %>%
  prcomp()

pcaCor <- pca$x %>%
  as.data.frame() %>%
  rownames_to_column("sample") %>%
  left_join(sumGc) %>%
  as_tibble() %>% 
  left_join(dgeList$samples) %>%
  dplyr::select(
    PC1, PC2, PC3, 
    Mean_GC = mn_gc, 
    Mean_Length = mn_logLen, 
    rRna_Proportion = rRNA
  ) %>% 
  cor()

a <- pca$x %>%
  as.data.frame() %>%
  rownames_to_column("sample") %>%
  left_join(dgeList$samples) %>%
  as_tibble() %>%
  ggplot(aes(PC1, PC2)) +
  geom_point(aes(colour = group), size = 2) +
  labs(
    x = paste0("PC1 (", percent(summary(pca)$importance["Proportion of Variance","PC1"]),")"),
    y = paste0("PC2 (", percent(summary(pca)$importance["Proportion of Variance","PC2"]),")"),
    colour = "Genotype"
  )
b <- pca$x %>%
  as.data.frame() %>%
  rownames_to_column("sample") %>%
  left_join(dgeList$samples) %>%
  ggplot(aes(PC1, rRNA, label = rRNA)) +
  geom_point(aes(colour = group), size = 2) +
  geom_smooth(method = "lm") +
  geom_text_repel(show.legend = FALSE) +
  scale_y_continuous(labels = percent) +
  labs(
    x = paste0("PC1 (", percent(summary(pca)$importance["Proportion of Variance","PC1"]),")"),
    y = "rRNA Proportion of Initial Library",
    colour = "Genotype"
  )
c <- pca$x %>%
  as.data.frame() %>%
  rownames_to_column("sample") %>%
  left_join(sumGc) %>%
  left_join(dgeList$samples) %>%
  as_tibble() %>%
  ggplot(aes(PC1, mn_logLen)) +
  geom_point(aes(colour = group), size = 2) +
  geom_smooth(method = "lm") +
  labs(
    x = paste0("PC1 (", percent(summary(pca)$importance["Proportion of Variance","PC1"]),")"),
    y = "Mean log(Length)",
    colour = "Genotype"
  )
d <- pca$x %>%
  as.data.frame() %>%
  rownames_to_column("sample") %>%
  left_join(sumGc) %>%
  left_join(dgeList$samples) %>%
  as_tibble() %>%
  ggplot(aes(PC1, mn_gc)) +
  geom_point(aes(colour = group), size = 2) +
  geom_smooth(method = "lm") +
  scale_y_continuous(labels = percent) +
  labs(
    x = paste0("PC1 (", percent(summary(pca)$importance["Proportion of Variance","PC1"]),")"),
    y = "Mean GC",
    colour = "Genotype"
  )
ggarrange(
  a, b, c, d, ncol = 2, nrow = 2,
  common.legend = TRUE, legend = "bottom"
) %>%
  annotate_figure("Psen2S4Ter")

PCA plot showing rRNA proportion, mean GC content and mean log(length) after summarisation to gene-level.

corrplot(
  pcaCor,
  type = "lower", 
  diag = FALSE, 
  addCoef.col = 1, addCoefasPercent = TRUE
)

Correlations between the first three principal components and measured variables: mean GC content, mean log(length) and rRNA proportion.

Differential expression

# dgeWTvHet <- dgeFilt[,str_detect(colnames(dgeFilt), "Het|WT")]

design <- model.matrix(~rRNA, data = dgeFilt$samples)

voom <- voom(dgeFilt, design = design)

fit <- lmFit(voom, design = design)

eBayes <- eBayes(fit)

topTable <- eBayes %>%
  topTable(coef = colnames(design)[2], sort.by = "p", n = Inf) %>%
  set_colnames(str_remove(colnames(.), "ID\\.")) %>%
  mutate(Bonf = p.adjust(P.Value, "bonferroni")) %>%
  mutate(DE = Bonf < 0.05) %>%
  unite(Location, c(seqnames, start, end, width, strand), sep = ":") %>%
  dplyr::select(
    Geneid = gene_id,
    Symbol = gene_name,
    AveExpr,
    logFC,
    P.Value,
    FDR = adj.P.Val,
    Location,
    t,
    DE,
    everything(),
    -B
  ) %>%
  as_tibble()

topTable %>% 
  dplyr::select(Geneid, Symbol, AveExpr, logFC, P.Value, FDR, DE) %>%
  mutate(
    AveExpr = format(round(AveExpr, 2), nsmall = 2),
    logFC = format(round(logFC, 2), nsmall = 2),
    P.Value = sprintf("%.2e", P.Value),
    FDR = sprintf("%.2e", FDR)
  ) %>%
  dplyr::slice(1:200) %>%
  datatable(
    options = list(pageLength = 20), 
    class = "striped hover condensed responsive", 
    filter = "top",
    caption = paste(
      "The top 100 differentially expressed genes.",
      nrow(dplyr::filter(topTable, DE)),
      "of",
      nrow(topTable),
      "genes were classified as DE with an Bonferroni p-value < 0.05."
    )
  )

Session info

save(
  addInfo,
  dgeList,
  topTable,
  file = here::here(
    "1_Psen2S4Ter/R/output/1_1_DE.RData"
  )
)

sessionInfo()

## R version 4.0.3 (2020-10-10)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 20.04.1 LTS
## 
## Matrix products: default
## BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.9.0
## LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.9.0
## 
## locale:
##  [1] LC_CTYPE=en_AU.UTF-8       LC_NUMERIC=C              
##  [3] LC_TIME=en_AU.UTF-8        LC_COLLATE=en_AU.UTF-8    
##  [5] LC_MONETARY=en_AU.UTF-8    LC_MESSAGES=en_AU.UTF-8   
##  [7] LC_PAPER=en_AU.UTF-8       LC_NAME=C                 
##  [9] LC_ADDRESS=C               LC_TELEPHONE=C            
## [11] LC_MEASUREMENT=en_AU.UTF-8 LC_IDENTIFICATION=C       
## 
## attached base packages:
## [1] stats4    parallel  stats     graphics  grDevices utils     datasets 
## [8] methods   base     
## 
## other attached packages:
##  [1] ggrepel_0.8.2           DT_0.14                 corrplot_0.84          
##  [4] edgeR_3.30.3            limma_3.44.3            ensembldb_2.12.1       
##  [7] AnnotationFilter_1.12.0 GenomicFeatures_1.40.1  AnnotationDbi_1.50.1   
## [10] Biobase_2.48.0          GenomicRanges_1.40.0    GenomeInfoDb_1.24.2    
## [13] IRanges_2.22.2          S4Vectors_0.26.1        AnnotationHub_2.20.0   
## [16] BiocFileCache_1.12.0    dbplyr_1.4.4            kableExtra_1.1.0       
## [19] ggpubr_0.4.0            scales_1.1.1            here_0.1               
## [22] ngsReports_1.4.2        BiocGenerics_0.34.0     magrittr_1.5           
## [25] forcats_0.5.0           stringr_1.4.0           dplyr_1.0.0            
## [28] purrr_0.3.4             readr_1.3.1             tidyr_1.1.0            
## [31] tibble_3.0.3            ggplot2_3.3.2           tidyverse_1.3.0        
## 
## loaded via a namespace (and not attached):
##   [1] readxl_1.3.1                  backports_1.1.8              
##   [3] plyr_1.8.6                    lazyeval_0.2.2               
##   [5] splines_4.0.3                 crosstalk_1.1.0.1            
##   [7] BiocParallel_1.22.0           digest_0.6.25                
##   [9] htmltools_0.5.0               fansi_0.4.1                  
##  [11] memoise_1.1.0                 cluster_2.1.0                
##  [13] openxlsx_4.1.5                Biostrings_2.56.0            
##  [15] modelr_0.1.8                  matrixStats_0.56.0           
##  [17] askpass_1.1                   prettyunits_1.1.1            
##  [19] jpeg_0.1-8.1                  colorspace_1.4-1             
##  [21] blob_1.2.1                    rvest_0.3.5                  
##  [23] rappdirs_0.3.1                haven_2.3.1                  
##  [25] xfun_0.15                     crayon_1.3.4                 
##  [27] RCurl_1.98-1.2                jsonlite_1.7.0               
##  [29] zoo_1.8-8                     glue_1.4.1                   
##  [31] gtable_0.3.0                  zlibbioc_1.34.0              
##  [33] XVector_0.28.0                webshot_0.5.2                
##  [35] DelayedArray_0.14.1           car_3.0-8                    
##  [37] abind_1.4-5                   DBI_1.1.0                    
##  [39] rstatix_0.6.0                 Rcpp_1.0.5                   
##  [41] progress_1.2.2                xtable_1.8-4                 
##  [43] viridisLite_0.3.0             flashClust_1.01-2            
##  [45] foreign_0.8-80                bit_1.1-15.2                 
##  [47] truncnorm_1.0-8               htmlwidgets_1.5.1            
##  [49] httr_1.4.1                    RColorBrewer_1.1-2           
##  [51] ellipsis_0.3.1                farver_2.0.3                 
##  [53] XML_3.99-0.4                  pkgconfig_2.0.3              
##  [55] locfit_1.5-9.4                labeling_0.3                 
##  [57] tidyselect_1.1.0              rlang_0.4.7                  
##  [59] reshape2_1.4.4                later_1.1.0.1                
##  [61] munsell_0.5.0                 BiocVersion_3.11.1           
##  [63] cellranger_1.1.0              tools_4.0.3                  
##  [65] cli_2.0.2                     generics_0.0.2               
##  [67] RSQLite_2.2.0                 broom_0.7.0                  
##  [69] fastmap_1.0.1                 evaluate_0.14                
##  [71] ggdendro_0.1.22               yaml_2.2.1                   
##  [73] knitr_1.29                    bit64_0.9-7.1                
##  [75] fs_1.4.2                      zip_2.0.4                    
##  [77] pander_0.6.3                  nlme_3.1-149                 
##  [79] mime_0.9                      leaps_3.1                    
##  [81] xml2_1.3.2                    biomaRt_2.44.1               
##  [83] compiler_4.0.3                rstudioapi_0.11              
##  [85] plotly_4.9.2.1                curl_4.3                     
##  [87] png_0.1-7                     interactiveDisplayBase_1.26.3
##  [89] ggsignif_0.6.0                reprex_0.3.0                 
##  [91] stringi_1.4.6                 highr_0.8                    
##  [93] lattice_0.20-41               ProtGenerics_1.20.0          
##  [95] Matrix_1.2-18                 vctrs_0.3.2                  
##  [97] pillar_1.4.6                  lifecycle_0.2.0              
##  [99] BiocManager_1.30.10           cowplot_1.0.0                
## [101] data.table_1.12.8             bitops_1.0-6                 
## [103] rtracklayer_1.48.0            httpuv_1.5.4                 
## [105] R6_2.4.1                      latticeExtra_0.6-29          
## [107] hwriter_1.3.2                 promises_1.1.1               
## [109] ShortRead_1.46.0              gridExtra_2.3                
## [111] rio_0.5.16                    MASS_7.3-53                  
## [113] assertthat_0.2.1              SummarizedExperiment_1.18.2  
## [115] openssl_1.4.2                 rprojroot_1.3-2              
## [117] withr_2.2.0                   GenomicAlignments_1.24.0     
## [119] Rsamtools_2.4.0               GenomeInfoDbData_1.2.3       
## [121] mgcv_1.8-33                   hms_0.5.3                    
## [123] grid_4.0.3                    rmarkdown_2.3                
## [125] carData_3.0-4                 Cairo_1.5-12.2               
## [127] scatterplot3d_0.3-41          shiny_1.5.0                  
## [129] lubridate_1.7.9               FactoMineR_2.3